Given the Internet of Things\u2019 (IoT) perch atop the hype cycle, IoT trend-spotting has become a full-time business, not just an end-of-the-year pastime. It seems every major \u2014 and minor \u2014 IoT player is busy laying out its vision of where the technology is going. Most of them harp on the same themes, of course, from massive growth to security vulnerabilities to skills shortages.\n\nThose are all real concerns, but Chris Nelson, vice president of engineering at operational intelligence (OT) vendor OSIsoft, shared some more unique viewpoints via email. In addition to his contention that the IoT will blur the lines between IT, which runs the customers\u2019 systems and email, and OT, which runs the technology behind the production systems, he talked about what will drive the IoT in the next year.\n8 trends driving the IoT in 2019\nLet\u2019s take a closer look at the eight trends Nelson thinking about:\n1. Industrial and commercial applications will drive the industry, not consumers\nAccording to Nelson, that\u2019s because businesses can monetize IoT\u2019s benefits better. He cites energy consumption as a key example, noting that \u201cindustry consumes 54 percent of delivered energy worldwide, according to the Energy Information Agency, or more than consumers or transportation combined.\u201d Reducing energy at an aluminum or paper plant by one or two percentage points, he says can mean millions of dollars in savings. A consumer cutting power consumption by 1 percent would save only a few dollars a month.\nMy take: True enough, but there are a lot more consumers than there are businesses, and overall consumer spending drives the U.S. economy. Still, Nelson has a point that relatively few big enterprise implementations could help jumpstart IoT usage, while energizing the mass consumer market can take years of expensive marketing to clarify sometimes complex and esoteric benefits. In addition, we can hope that industrial and enterprise IoT will be better equipped to deal with security concerns.\n2. The edge will be far more important than people realize\n\u201cThe edge is basically any place \u2014 a wind farm, a factory \u2014 where data is generated, analyzed, and largely stored locally,\u201d Nelson said. \u201cWait? Isn\u2019t that just a data center? Sort of. The difference is the Internet of Things.\u201d His point is that most of the vast amounts of data that is machine-generated doesn\u2019t need to go very far. \u201cThe people who want it and use it are generally in the same building,\u201d he noted, quoting Gartner\u2019s prediction that more than 50 percent of data will be generated and processed outside traditional data centers \u2014 on the edge \u2014 although \u201csnapshots and summaries might go to the cloud for deep analytics.\u201d\nBut Nelson wasn\u2019t sure about what kind of edge architectures would prevail. The edge might function like an interim way station for the cloud, he noted, or we could see the emergence of \u201cZone\u201d networking \u2014 edges within edges \u2014 that can conduct their own analytics and perform other tasks on a smaller, more efficient scale.\nMy take: It\u2019s hard to disagree with the synergy between the IoT and \u201cthe edge,\u201d but as the cloud architectures continue to dominate computing architectures, it seems like the architectural distinctions among the edge, the data center, and the cloud may start to fade.\n3. Synthetic data will become a more urgent concern\nNelson defines synthetic data as \u201cmisleading information that makes good people do bad things.\u201d He means things like hackers sending \u201csynthetic\u201d notifications to a control room to get operators to open gates on a reservoir, flooding a neighborhood. He calls it \u201cStuxnet goes mainstream.\u201d He noted efforts by Lawrence Berkeley Lab and Aperio, among others, on various efforts to spot fake data.\nMy take: Given my work in the application monitoring space, "synthetic data" means something a bit different to me. But Nelson is right that hacked or spoofed IoT data is a currently under-appreciated risk.\n4. Real-time data will grow in importance\nNelson cited IDC data that says\u00a0real-time data will grow from 15 percent of digital data in 2017 to 30 percent in 2030, with a jump of 7x to 10 in total data volume. He predicts more innovation and investment in this area, \u201cparticularly in software that will let people understand what machines are saying.\u201d\nMy take: There\u2019s no question that using real-time data to drive real-time decisions will become increasingly important. Given the huge amounts of data involved, though, I would look for artificial intelligence (AI) and machine learning solutions to take the forefront in turning this data into action.\n5. Smart equipment will begin to get momentum\nNelson predicted that \u201cmanufacturers will increasingly integrate real-time monitoring and diagnostics into equipment,\u201d using Caterpillar\u2019s CAT Connect engine-monitoring services and Flowserve\u2019s building intelligence services into industrial pumps.\n\u201cOver the past five years,\u201d he said, \u201cwe\u2019ve seen the technology stack come together and several end-users conduct trials. Over the next five, we will see commercial adoption.\u201d\nMy take: Again, I have no argument with this point, but building smarts into expensive, long-lasting industrial equipment carries its own risks in the fast-evolving world of IoT. Because IoT changes much faster than the useful life of the equipment, there\u2019s a real risk of obsolescence unless vendors can create modular, upgradable, solutions.\n\n6. Rules and business practices for data sharing will start to gel\nNelson posed an interesting question: \u201cLet\u2019s say an equipment provider provides ongoing monitoring on devices it sold or leased to an end user. Who owns that data? Most would say the end users, but what if the equipment provider conducted analytics on the raw data thereby creating a second set of information that\u2019s more valuable than the first? Can data from one facility be anonymized and used to optimize benchmarks for another owned by a competitor? These are big questions, and no one has figured them out yet.\nMy take: Yes, yes, yes. But data ownership is a huge and thorny issue, and I am less confident that we\u2019ll make real progress on solving it in 2019 or any time soon. I look for this to be an ongoing area of concern for years.\n7. Traditional businesses will develop new business models out of IoT\nNelson cited small, rural utilities that have begun to sell broadband services by leveraging their investments for smart meters in a new way, as well as \u201clarge utilities and manufacturers study plans to commercialize their in-house IoT applications for predictive maintenance.\u201d\nMy take: That\u2019s only the tip of the iceberg for what many see as the holy grail of IoT. Sure, saving money is great, but the real opportunity is using IoT to create wholly new businesses. I think it\u2019s still too early to know what those new ideas will be and which ones will take off.\n8. IoT projects will have to hit their numbers\n\u201cCompanies won\u2019t fund open-ended projects,\u201d Nelson said, and \u201cthey will want to see payoff in two years or less.\u201d\nMy take: Not sure we\u2019re there yet. Many enterprise and industrial IoT projects are still in the pilot stage, trying to figure out what their \u201cnumbers\u201d should be. Until that gets settled, it seems a little premature to talk about \u201cmaking\u201d those numbers.